materials
Article
Additively Manufactured Composite Lug with Continuous
Carbon Fibre Steering Based on Finite Element Analysis
Chethan Savandaiah 1 , Stefan Sieberer 2, * and Georg Steinbichler 3






Citation: Savandaiah, C.; Sieberer, S.;
Steinbichler, G. Additively
Manufactured Composite Lug with
Continuous Carbon Fibre Steering
Based on Finite Element Analysis.
Materials 2022, 15, 1820. https://
doi.org/10.3390/ma15051820
Academic Editors:
Andrea Bernasconi and Luca
Michele Martulli
Received: 8 February 2022
Accepted: 25 February 2022
Published: 28 February 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Materials 2022, 15, 1820. https://doi.org/10.3390/ma15051820
1 Biobased Composite and Processes, Wood K Plus-Kompetenzzentrum Holz GmbH, 4040 Linz, Austria;
c.savandaiah@wood-kplus.at
2 Institute of Structural Lightweight Design, Johannes Kepler University Linz, 4040 Linz, Austria
3 Institute of Injection Moulding and Process Automation, Johannes Kepler University Linz, 4040 Linz, Austria;
georg.steinbichler@jku.at
* Correspondence: stefan.sieberer@jku.at
Abstract: In this study, the influence of curvilinear fibre reinforcement on the load-carrying capac-
ity of additively manufactured continuous carbon fibre reinforced necked double shear lugs was
investigated. A curvilinear fibre placement is descriptive of layers in extrusion-based continuous-
fibre-reinforced additive manufacturing with carbon fibres aligned in the directions of principal
stress. The alternating layered fibre trajectories follow the maximum and minimum principal stress
directions due to axial tension loading derived from two-dimensional finite element analysis (FEA).
The digital image correlation was utilised to monitor the strain distribution during the application
of tensile load. The 2D FEA data and the tensile test results obtained were comparable, the part
strength and the linear approximation of stiffness data variability were minimal and well within the
acceptable range. Nondestructive fractography was performed by utilising computed tomography
(CT) to analyse the fractured regions of the tensile-tested lug. The CT scanned images aided in
deducing the failure phenomenon in layered lugs; process-induced voids and fibre layup undulation
were identified as the cause for lug failure.
Keywords: additive manufacturing; continuous fibre reinforcement; principal strain; optimisation;
stiffness and strength analysis; experimental testing; digital image correlation
1. Introduction
The behaviour of materials extrusion (MEX) additive manufacturing of short-fibre-
reinforced thermoplastic composites had been well studied as processing effects on ther-
momechanical and morphological properties [1,2]. Several researchers have investigated
the influence of fibre length distribution [3,4], process-induced anisotropy [5] and thermal
material [6] properties on the functional performance [7] of the parts produced via MEX
composite additive manufacturing. The key aspect for the adoption of composite MEX is
the light-weight potential of parts and assemblies by functional optimisation of the end-user
application [8]. The design considerations for MEX composite structures are based on the
multiple layers of fibre-reinforced matrix [9]. The short fibres in each layer are straight
and parallel to each other and mostly aligned in the printing direction, causing anisotropic
material behaviour [10,11]. Optimising the fibre alignment within each layer to alternating
orientation from point to point has not been adopted, albeit due to the nonavailability of
necessary commercial tools. Moreover, MEX composite parts are fairly lucrative in terms of
overall economics but have poor mechanical and thermal material characteristics [12,13].
The innovative continuous carbon fibre reinforced composite materials in MEX have
yet to realise full potential [14,15]. Azarov et al. [16] assessed the complete product devel-
opment cycle, determining the mechanical properties for the frame material, modelling
and the structural analysis of small-scaled autonomous drones. The authors concluded that
MEX type processing of continuous fibre composites is a highly promising technology for
https://www.mdpi.com/journal/materials
Materials 2022, 15, 1820 2 of 14

the manufacturing of lightweight structures for recreational and industrial drones alike.
Furthermore, Borowski et al. [17] studied process-induced consolidation of continuous
carbon fibre polycarbonate, including the parameterisation of printing temperature to print
standard coupons, and performed three-point bending tests to determine the quality of
the in situ consolidated MEX coupons. Furthermore, a highly curved test structure was
fabricated to determine the limits of the processing with help of computed tomography.
Similarly, Dickson et al. [18] evaluated the performance of several MEX-processed carbon,
aramid and glass continuous fibre reinforced standard coupons subjected to tension and
flexure standard testing. They concluded that the tensile strength of continuous carbon
fibre reinforced MEX coupons was up to 6.3 times that of nonreinforced thermoplastic
MEX coupons.
Zhu et al. [19] proposed fibre path optimisation to improve the uniaxial tensile load
carrying ability of composite laminates with a centred hole using finite element method.
The researchers reported a reduction in maximum strength indexes ranging between 19%
based on maximum stress criterion and 39% based on Tsai–Wu failure criterion. In addition,
Sugiyama et al. [20] and Ferreira et al. [21] proposed optimised variable fibre volume frac-
tion and stiffness composites for curved fibre trajectory along the principal stress direction
around the hole. The specimens fabricated with optimised fibre trajectories using continu-
ous carbon fibre additive manufacturing were 1.6 times stronger than conventional linear
laminate composites. The use of carbon-fibre-reinforced composite materials in this manner
omits its major attribution: optimised strength and stiffness may be aligned in the direc-
tions which correspond to the applied loads. In terms of optimisation, Ghiasi et al. [22,23]
suggest two distinct design considerations, constant stiffness and variable stiffness. The
constant stiffness design terminology in MEX composite additive manufacturing relates to
fibre layup with a similar raster angle for the whole fabricated part. However, in variable
stiffness design, the optimised fibre paths reinforce the high strain regions, thus eliminating
the excess use of fibre reinforcement and simultaneously decreasing the part weight.
The structural design under investigation in this study, a necked double shear lug,
contains a geometric discontinuity, a hole, which interposes the fibre continuity and causes
a concentration and realignment of stress. The specific issue in this study is to use fibre
reinforcement in such a way that the direction of the fibres, or at least some of the fibres, is a
function of spatial position in the structure which may lead to steep increases in structural
performance. This particular study examines the departure from constant stiffness design
use of composite materials in additive manufacturing. The researchers investigated a
variable stiffness design approach for the structural part based on principal stress directions
fabricated via MEX composite additive manufacturing. In this study, within the variable
stiffness design, an optimised elementwise design is considered [24,25]. It creates a distinct
orientation for every finite element of a mesh representing a necked double shear lug based
on the simulated load conditions. Furthermore, it is postprocessed to derive the final fibre
paths for feasible processing in MEX.

2. Materials and Methodology

A commercially available 20 wt.% short carbon fibre (diameter 7 µm, measured
weighted average length 220 µm and density 1.72 g·cm−3 ) reinforced polyamide 6 (PA6)
filament with a diameter of 1.75 mm was supplied by Prirevo 3D Solutions GmbH, Ried
im Traunkreis, Austria, and the material data were taken from the article (see Figure S1).
Additionally, the PA6 infused continuous carbon fibre (CCF) with 0.36 mm diameter was
purchased from Markforged, Watertown, MA, USA, with fibre volume fraction of approxi-
mately 35% [26].
To orientate the fibres best to the principal strain or stress directions in the considered
part with fixed geometrical boundary conditions, a numerical approach was applied. First,
an isotropic 2D finite element (FE) model was set up in ABAQUS (2017, 3DS, Waltham, MA,
USA), assuming quasi-isotropic, linear elastic material behaviour [8]. Quadratic elements
are used as they are necessary for the bending part, and nonlinear geometry is used for
Materials 2022, 15, 1820 3 of 14

the stress and strain calculation in FE. Laminate stiffness values are taken from eLamX
software (V2.6, TU Dresden, Germany) using 0◦ stiffness of the CCF (80000 MPa, value
taken from in-house test setup) and assuming transversal and shear values. For this initial
model, the directions of principal stresses and strains under the presumed test load case
were obtained. These data were postprocessed in MATLAB (R2017, MathWorks, Natick,
MA, USA) where principal strain orientations were calculated for each element based on
the isotropic results (see, e.g., [27] for further details on strain trajectory calculation and
potential applications). Principal axis orientations were then fed back to the FE model as
individual element material orientations. Using postprocessed data points from MATLAB,
alternating layers of minimum and maximum principal stress fibre trajectories were traced
within the outline of the part with help of spline command in computer-aided drafting
software NanoCAD 5.0. (Nanosoft, Moscow, Russia) A proprietary plugin developed
for the drafting software, NanoGcode (v. 0.3.4.41535, Nanosoft, Moscow, Russia), was
provided by Anisoprint and was utilised to generate print settings and subsequent G-codes
for printing curvilinear fibre paths in Anisoprint A4 Composer 3D printer.
The details of print and layer settings are summarised in Table 1, and the specimen
dimensions are shown in Figure 1. A slight modification to the existing A4 Composer
printing head was required to print the thermoplastic-based CCF. The printer consists
of two print heads, a conventional print head for printing pure thermoplastic or filled
thermoplastic filament and a composite coextrusion print head, which is analogous to cable
coating. The composite coextrusion print head is designed to print proprietary thermoset-
based CCF via coating with thermoplastic (as a binder). Hence, a cylindrical stainless steel
tube with an internal diameter of 2 mm inlay with polytetrafluorethylene microtubing with
an internal diameter of 0.45 mm was used as a thermoplastic CCF feeding tube (Figure 2).
The feeding tube acts as a thermal barrier until the nozzle exit, and upon exiting the nozzle,
the fibres are uniformly ironed onto the previous layer with help of a tapered nozzle (see
Figure 2 inset).

Table 1. Detailed printing settings for CCF and FFF print head.

Print Settings Unit Value

Layer height mm 0.15
CCF layer width mm 1
FFF layer width mm 0.2
Nozzle diameter mm 0.7
Lug thickness mm 8.5
CCF print head temperature ◦C 250
FFF print head temperature ◦C 260
Print bed temperature ◦C 95
CCF print speed (red) mm·min−1 550
CCF print speed (green) mm·min−1 400
FFF print speed mm·min−1 2200
Maximum principal stress layers (CCF and FFF print head) - 29
Minimum principal stress layers (CCF and FFF print head) - 27
Top plastic layer (FFF print head) - 1
Materials 2022, 15,
Materials 2022, 15, 1820
x FOR PEER REVIEW 44 of
of 14
15
Materials 2022, 15, x FOR PEER REVIEW 4 of 15

Figure 1.
1. Double sheared
sheared lug specimen
specimen dimensions.(mm).
Figure 1. Double
Figure Double sheared lug
lug specimen dimensions
dimensions.

Figure 2. Modified CFC print head from Anisoprint for printing CCF; nozzle in the inset.
Figure 2.
Figure 2. Modified
Modified CFC
CFC print
print head
head from
from Anisoprint
Anisoprintfor
forprinting
printingCCF;
CCF;nozzle
nozzlein
inthe
theinset.
inset.

Optimised manufacturing parameters for printing with CCF were utilised to

Optimisedmanufacturing
Optimised manufacturing parameters
parameters for printing
for printing with CCF withwere
CCFutilised
were toutilised
minimise to
minimise the process-induced defects. The different coloured CCF layup in Figure 3b
minimise
the the process-induced
process-induced defects. The defects. Thecoloured
different differentCCFcoloured
layupCCF layup3b
in Figure inrepresents
Figure 3b
represents transient printing speed to accommodate the printing accuracy around sharp
represents
transient transient
printing printing
speed speed to accommodate
to accommodate the printingthe printingaround
accuracy accuracy around
sharp turns;sharp
red
turns; red colour represents printing at a higher speed compared to the green coloured
turns; red colour represents printing at a higher speed compared to
colour represents printing at a higher speed compared to the green coloured CCF layup (seethe green coloured
CCF layup (see Figure 3). Moreover, the large gaps observed between the CCF layup in
CCF layup
Table (see Figure
1). Moreover, 3). Moreover,
the large the large
gaps observed gapsthe
between observed
CCF layup between the CCF
in Figure layup
3b,c are in
filled
Figure 3b,c are filled with thermoplastic infills to reduce the voids. Furthermore, the CCF
Figure
with 3b,c are filled
thermoplastic withtothermoplastic
infills infillsFurthermore,
reduce the voids. to reduce thethe voids.
CCFFurthermore, the CCF
layup is a continuous
layup is a continuous loop, meaning each layer has just one fibre cut at the end of the CCF
layupmeaning
loop, is a continuous loop,has
each layer meaning each
just one fibrelayer
cuthas justend
at the oneof fibre
thecut
CCF at printing,
the end ofenabling
the CCF
printing, enabling fibre layup accuracy and reducing the overall print time. The maximum
printing,
fibre layup enabling
accuracy fibre
and layup accuracy
reducing and reducing
the overall the overall
print time. print time.principal
The maximum The maximumstrain
principal strain trajectories are shown in red and green (Figure 3b). Minimum principal
principal strain
trajectories trajectories
are shown in redare
andshown in red and
green (Figure 3b).green
Minimum(Figure 3b). Minimum
principal principal
strain trajectories
strain
are trajectories are
perpendicular perpendicular
to the to the maximum
maximum principal strains andprincipal strains and
are exemplarily are exemplarily
shown in green in
strain trajectories are perpendicular to the maximum principal strains and are exemplarily
shown
the shaftinregion
green in thethe shaft region3c);
of the lug (Figure 3c); the dark grey coloured section
shown in green ofin the lug
shaft(Figure
region of the dark grey coloured
lug (Figure 3c); the section
dark greycorresponds to just
coloured section
corresponds toprinting
thermoplastic just thermoplastic
without CCF. printing without CCF.
corresponds to just thermoplastic printing without CCF.
Materials 2022, 15, 1820 5 of 14
Materials 2022, 15, x FOR PEER REVIEW 5 of 15

Figure 3. Principal strain trajectories based on the isotropic FE model; green and red coloured
Figure 3. Principal strain trajectories based on the isotropic FE model; green and red coloured
splines represent CCF layup, and grey and dark grey colours represent thermoplastic for perimeters
splines represent
and infills. Top CCF layup, and
and bottom grey and dark
thermoplastic layergrey colours represent
(a); maximum principalthermoplastic forCCF
stress direction perimeters
(b),
and infills. Top
minimum and bottom
principal thermoplastic
stress direction CCF (c).layer (a); maximum principal stress direction CCF (b),
minimum principal stress direction CCF (c).
From the initial results, the print layup was obtained and an updated FE model with
From thematerial
orthotropic initial results, the print
properties waslayup wasFigure
set up. obtained and an updated
4 shows the meshFEandmodel with or-
material
thotropic material
orientations usedproperties was set up.
in the orthotropic FE Figure
model. 4The
shows the mesh
outside and material
perimeter (red) hasorientations
no CCF
used in the orthotropic
reinforcements, FE model.
the region around Thethe outside perimeter
lug head (blue) (red) has no CCF-reinforced
is uniaxial CCF reinforcements, in
the region around
maximum principalthe lug head
strain (blue)
direction is uniaxial
according to theCCF-reinforced
previous results,inand
maximum principal
the shaft (beige)
strain direction
section according
is biaxial to the previous
CCF-reinforced, in bothresults, andstrain
principal the shaft (beige)This
directions. section is biaxial
modelling
reflects the printing
CCF-reinforced, in bothrestrictions
principalinstrain
the lug head andThis
directions. the modelling
necessary perimeter.
reflects theThe FE
printing
analysis was
restrictions repeated,
in the lug headand and
stresstheand strain data
necessary for the part
perimeter. The with optimal was
FE analysis CCFrepeated,
layup
were
and yielded.
stress Moreover,
and strain data the
for curvilinear
the part with layup of theCCF
optimal specimen
layupcorresponds
were yielded.largely but
Moreover,
not fully with the material editing in the FE model; however, this discrepancy
the curvilinear layup of the specimen corresponds largely but not fully with the material is small.
editing in the FE model; however, this discrepancy is small.
A total of four specimens were printed, and the experimental testing was performed
on a servohydraulic test rig (Zwick Roell, Ulm, Germany) with a cylinder rated at 100 kN
force. The shaft end of the specimen was designed with a flat, wide end (see Figure 1)
to be clamped in hydraulic wedge grips. A purpose-built test jig using a steel pin with
23 mm diameter was used to fix the lug eye to the test bed. The test rig was operated
by Cubus software (V2.48, Zwick Roell, Ulm, Germany) in displacement mode with a
constant rate of 1 mm/min. A Zwick Roell force transducer and the internal displacement-
sensor of the cylinder were used for measurement of force and displacement, respectively.
Additionally, a Correlated Solutions 3D digital image correlation (DIC) system recorded
the surface displacement and calculated the surface strains of the specimens during the
tests. Postprocessing of the DIC data was performed in the software Vic 3D 8 (Correlated
Solutions, Irmo, SC, USA).
Materials 2022, 15,
Materials 2022, 15, 1820
x FOR PEER REVIEW 66 of
of 14
15

Figure 4. Material
Figure 4. Material and
and mesh
mesh on
on the
the orthotropic
orthotropic model.
model.

X-ray computed tomography (CT) is used for characterising the fractured specimens.
A total of four specimens were printed, and the experimental testing was performed
Therefore, scans were conducted on a Nanotom 180 NF (GE Phoenix X-ray, Niskayuna, NY,
on a servohydraulic test rig (Zwick Roell, Ulm, Germany) with a cylinder rated at 100 kN
USA) laboratory CT device. A molybdenum target and a tube voltage of 60 kV were used
force. The shaft end of the specimen was designed with a flat, wide end (see Figure 1) to
for the acquisition of the data.
be clamped in hydraulic wedge grips. A purpose-built test jig using a steel pin with 23
A scanning electron microscope (SEM) was used for fractography of CCF filament on
mm diameter was used to fix the lug eye to the test bed. The test rig was operated by
Phenom Pro X (Thermo Fisher Scientific, Waltham, MA, USA), using secondary electrons
Cubus software (V2.48, Zwick Roell, Ulm, Germany) in displacement mode with a
at 15 kV.
constant rate of 1 mm/min. A Zwick Roell force transducer and the internal
displacement-sensor
3. Result and Discussion of the cylinder were used for measurement of force and
displacement,
3.1. Fractography respectively. Additionally, a Correlated Solutions 3D digital image
correlation (DIC) system recorded the surface displacement and calculated the surface
A CT scan of fractured lug specimen and cryofractured CCF was performed to in-
strains of the specimens during the tests. Postprocessing of the DIC data was performed
vestigate the fracture behaviour of the additively manufactured layered composite and to
in the software Vic 3D 8 (Correlated Solutions, Irmo, SC, USA).
identify the inherent defects within the CCF. In Figure 5a, analysis of the CT image of CCF
X-ray computed tomography (CT) is used for characterising the fractured specimens.
assists in identifying the bright region as carbon fibre bundles, light greyscale region as
Therefore, scans were conducted on a Nanotom 180 NF (GE Phoenix X-ray, Niskayuna,
the thermoplastic zone and dark spots highlighted within red circles as voids. Similarly,
NY, USA)
Figure 5b islaboratory CT device.
a CT sectional image A
of molybdenum
the CCF with target and
sections of avoids
tubehighlighted
voltage of 60 kV were
within the
used for the acquisition of the data.
red capsules. The quantifiable void content within the CCF was found to be approximately
 A CT scan of fractured lug specimen and cryofractured CCF was performed to
investigate the fracture behaviour of the additively manufactured layered composite and
to identify the inherent defects within the CCF. In Figure 5a, analysis of the CT image of
CCF assists in identifying the bright region as carbon fibre bundles, light greyscale region
as the thermoplastic zone and dark spots highlighted within red circles as voids. Similarly,
Materials 2022, 15, 1820 7 of 14
Figure 5b is a CT sectional image of the CCF with sections of voids highlighted within the
red capsules. The quantifiable void content within the CCF was found to be
approximately 1.5 vol.%. In Figure 5c, the SEM image of cryofractured CCF validates the
1.5 vol.%. In
detection of Figure 5c, thethe
voids within SEM image
CCF in CTof scans.
cryofractured
The loose CCF validates
carbon fibresthe detection
(red box andof voids
circle)
within
observed are due to low infusion of highly viscous PA6 thermoplastic matrix betweento
the CCF in CT scans. The loose carbon fibres (red box and circle) observed are due
low infusionfibres
individual of highly viscous
causing fibrePA6 thermoplastic
slippage matrix air
and entrapping between
bubbles individual
within thefibres
CCF. causing
This
fibre
is detrimental to the overall performance of CCF, as the capability of the carbonoverall
slippage and entrapping air bubbles within the CCF. This is detrimental to the fibre
performance
bundle to carry of CCF,
strainasuniformly
the capability of the carbon
is diminished. fibre
Single bundle
carbon to carry
fibre strain
fracture uniformly
results in earlyis
diminished.
unpredictable composite failures, as highlighted in the inset of Figure 6a (red box).as
Single carbon fibre fracture results in early unpredictable composite failures,
highlighted
Furthermore, in the
the process-induced
inset of Figure 6avoids
(red box). Furthermore,
are also a cause forthe process-induced
drastic voids are
influence in forming
also a cause in
undulation forCCF
drastic influence
layup shown in in forming
Figure 6aundulation
(red box), in CCF as
mainly layup shown
a result in Figureof6a
of buckling
(red
loosebox), mainly
carbon as a during
fibres result ofCCF
buckling of loose
extrusion andcarbon fibres
in situ during CCF extrusion
consolidation and in
due to nozzle
situ consolidation due to nozzle compaction, similarly
compaction, similarly reviewed by Sanei and Popescu [28]. reviewed by Sanei and Popescu [28].

Figure 5. Images acquired via CT scan and SEM. Overview of CCF via CT (a); sectional view of CCF
Figure 5. Images acquired via CT scan and SEM. Overview of CCF via CT (a); sectional view of CCF
via CT (b); cryofractured surface of CCF via SEM (c).
via CT (b); cryofractured surface of CCF via SEM (c).
Figure6a
Figure 6ashows
shows a single
a single layer
layer maximum
maximum principal
principal direction
direction layuplayup
along along
with awith
layereda
layered composite primary fracture zone CT image as an inset.
composite primary fracture zone CT image as an inset. Figure 6b shows the overview and Figure 6b shows the
overviewand
primary andsecondary
primary and secondary
fractures fractures
as the side viewas the sidetensile-tested
of the view of the tensile-tested
lug. The primarylug.
The primary fracture is indicated by the red frame, and the secondary
fracture is indicated by the red frame, and the secondary fracture is represented within fracture is
represented within the yellow frame. The CCF layers break within
the yellow frame. The CCF layers break within a narrow window around the location of a narrow window
around
the the location
maximum stress,ofindicating
the maximum somestress,
strengthindicating
variationsome
overstrength
the lengthvariation
of the over
CCF theand
length of the CCF and potentially undulation-induced load variation
potentially undulation-induced load variation between layers highlighted in Figure 6a (red between layers
highlighted
box), and theinreasoning
Figure 6ais(red
thatbox), anddiscussed
already the reasoning is that5.already
in Figure In Figure discussed
6b, the CTin Figure
scan of
both primary and secondary fractures shows uneven failure, which is an attributefailure,
5. In Figure 6b, the CT scan of both primary and secondary fractures shows uneven of MEX
which is ansimilarly
specimens attributereported
of MEX by specimens
authors similarly
Savandaiah reported byand
et al. [3] authors
SpoerkSavandaiah
et al. [5]. et al.
Voids
[3] and Spoerk et al. [5]. Voids between each layer are inherent to
between each layer are inherent to MEX due to layered processing and result in poor loadMEX due to layered
transfer between layers, causing jagged teeth fractured appearance [4,17]. Furthermore, in
Figure 6b side views, the CT scan shows some delamination, which may have occurred
during fracture of single layers; however, this cannot be resolved from the postfracture
examination data. In Figure 6c, the minimum principal direction layup and the twisted
CCF bundles around the curved sections are shown. According to the authors Shiratori
et al. [29], twisting due to curving caused the early damage to fibres during the CCF
layup, and the process-induced voids within the broken and twisted CCF may negatively
influence the overall mechanical performance of the MEX specimens [30]. Videos of the CT
scan data are provided in the supplementary data in Videos S1 and S2 for front and side
views, respectively.
 delamination, which may have occurred during fracture of single layers; however, this
cannot be resolved from the postfracture examination data. In Figure 6c, the minimum
principal direction layup and the twisted CCF bundles around the curved sections are
shown. According to the authors Shiratori et al. [29], twisting due to curving caused the
early damage to fibres during the CCF layup, and the process-induced voids within the
Materials 2022, 15, 1820 8 of 14
broken and twisted CCF may negatively influence the overall mechanical performance of
the MEX specimens [30].

Figure 6. CT scan of the fractured lug with an overview of two fractured zones. Maximum principal
Figure 6. CT scan of the fractured lug with an overview of two fractured zones. Maximum principal
stress direction CCF layup with a primary fractured zone as an inset (a); overview and side views
stress direction
of primary (redCCF layup
frame) with
and a primary
secondary fractured
(yellow zonefractures
frame) as an inset (a);minimum
(b); overview principal
and side views
stress
of primary (red frame) and secondary (yellow frame) fractures (b); minimum principal stressregion
direction CCF layup (c). The red connecting line indicates the cause and effect in the fractured direc-
tion CCF layup (c).
(undulation and voids). The red connecting line indicates the cause and effect in the fractured region
(undulation and voids).
3.2. FEA and Strain Trajectory Analysis
3.2. FEA and Strain Trajectory Analysis
Figure 7 gives strain results for the plane-strain model with a tensile line load of 1300
N·mm Figure 7 gives
−1 applied strain
to the shaftresults for which
of the lug the plane-strain
is reacted inmodel
the lugwith
eye. aThis
tensile
loadline
wasload of
chosen
1300 N · mm −1 applied to the shaft of the lug which is reacted in the lug eye. This load was
for comparison to measurement data later in this paper. The maximum principal strain in
chosen
the neckforregion
comparison to measurement
is 1,max,neck data
= 8.10·e−3, and lateroutermost
at the in this paper.
pointThe
formaximum
which theprincipal
DIC can
strain in the neck region is ε1,max,neck = 8.10 × 10 −3 , and at the outermost point for
which the DIC can obtain values, 1 mm inside the edge for this analysis, the strain is
ε1,DIC,neck = 7.23 × 10−3 . These strains are in the perimeter region where only short-fibre-
reinforced thermoplastic material was printed. The absolute maximum of principal strains
is ε1,max = 8.50 × 10−3 at the lug eye, where CCF was printed. Because of the modulus
difference between thermoplastic and CCF material, the stresses are much larger at the lug
eye than at the neck. However, the area around the lug eye was obstructed from view in
testing by the test fixture, and no strain measurement was taken here. In the shaft (top),
away from the strain concentrations, a quasi-uniaxial strain of ε1,shaft = 2.35 × 10−3 was
 strains are in the perimeter region where only short-fibre-reinforced thermoplastic
material was printed. The absolute maximum of principal strains is 1,max = 8.50·e−3 at the
lug eye, where CCF was printed. Because of the modulus difference between
thermoplastic and CCF material, the stresses are much larger at the lug eye than at the
Materials 2022, 15, 1820 neck. However, the area around the lug eye was obstructed from view in testing by9 of the
14
test fixture, and no strain measurement was taken here. In the shaft (top), away from the
strain concentrations, a quasi-uniaxial strain of 1,shaft = 2.35·e−3 was obtained. The values
are compared
obtained. Theto measured
values strains into
are compared Table 2. The strains
measured quasi-isotropic
in Table (see
2. TheFigure 8) and the
quasi-isotropic
curvilinear
(see Figure 8) and the curvilinear orthotropic designs were all predicted to fail due the
orthotropic designs were all predicted to fail due to fibre tension near lug
to fibre
eye. Though the curvilinear design had the same failure mode, it achieved
tension near the lug eye. Though the curvilinear design had the same failure mode, it the design
consideration of transferring
achieved the design the applied
consideration load effectively
of transferring around
the applied loadthe eye. The around
effectively strain near
the
the
eye. The strain near the lug eye was highest in quasi-isotropic design, in contrast tonear
lug eye was highest in quasi-isotropic design, in contrast to the material the
curvilinear
material neardesign.
curvilinear design.

Figure
Figure 7.
7. Maximum
Maximum principal
principal strains
strains on
on the
the orthotropic
orthotropic model
model under
under tensile
tensile load.
load. Full
Full lug
lug (a);
(a);
detailed neck area where the DIC evaluation area limit is indicated (b).
detailed neck area where the DIC evaluation area limit is indicated (b).
Table 2. Comparison of FE and experimental test results: maximum strain around the neck
Table 2. Comparison of FE and experimental test results: maximum strain around the neck compared
compared to strain at shaft.
to strain at shaft.
Position FE DIC
Position FE DIC
Shaft (1,shaft) 0.00235 0.00245
Shaft (ε
Neck (1,DIC,neck ) 1,shaft ) 0.00235
0.00723 0.0073 (avg. left and0.00245
right)
Neck (ε1,DIC,neck ) 0.00723 0.0073 (avg. left and right)
Strain concentration factor
Strain concentration factor
3.08 3.08 3.02 3.02
Materials 2022,
Materials 15, 15,
2022, 1820x FOR PEER REVIEW 10 of1015of 14

Figure 8. Maximum principal strains for unit tensile load on the lugs. Short carbon fibre reinforced
Figure 8. Maximum principal strains for unit tensile load on the lugs. Short carbon fibre reinforced
PA6 quasi-isotropic material (a); load-path curvilinear continuous carbon fibre reinforced material
PA6
(b).quasi-isotropic
The significant material
reduction(a);
in load-path
strain due curvilinear continuous carbon
to the fibre reinforcement fibrevisible.
is clearly reinforced material (b).
Additionally,
The
thesignificant reduction
superior strength in CCF
of the strainmaterial
due to may
the fibre reinforcement
be fully utilised. is clearly visible. Additionally, the
superior strength of the CCF material may be fully utilised.
3.3. Evaluation of Modelling by Experimental Testing
3.3. Evaluation of Modelling by Experimental Testing
In Figure 9, the image clearly shows the loading conditions of the tested lug and the
In Figure
strain 9, theThe
variations. image clearlyFE
obtained shows
resultsthewere
loading conditions
compared of the
against tested lug and
experimental testthe
strain variations.
results, in which The obtained
surface strainsFE results
were were compared
monitored by optical against experimental
DIC measurement. test results,
Because of
inthe
which
mounting surface strains
of the lug, not were monitored
all of the surface bywas
optical
visibleDIC measurement.
to the DIC. However, Because of the
high strain
mounting
areas at the of the
neck lug,
of not
the all
lugofwere
the surface
clearly was visible
in view. Thetotest
the specimen
DIC. However,
lug No.high strain areas
2 provided
atexcellent
the neckDIC of the lug were
patterns andclearly
insightin into
view.material
The testbehaviour.
specimen Figure
lug No.9 2shows
provided
the excellent
major
DIC patterns
principal andobtained
strain insight into material
in testing at a behaviour.
load of 11,590Figure 9 shows
N. The the majorallows
DIC evaluation principal strain
strains
obtained in testing
to be measured at aa load
only finiteofdistance
11,590 N. fromThethe
DIC evaluation
part boundary,allows
and instrains to be
the case measured
of the lug
only
this aisfinite
aboutdistance
1 mm. Note from the the similarity
part boundary,in theand in the
strain case of the
distribution lug this to
compared is about
Figure17.mm.
This strain pattern also makes specimens more susceptible to failures
Note the similarity in the strain distribution compared to Figure 7. This strain pattern due to MEX defectsalso
reported
makes earlier more
specimens causing premature
susceptible to failure
failuresofdue theto specimen
MEX defects duereported
to the strain
earlierbeing
causing
concentrated
premature in the
failure of individual
the specimen CCFdue layers. This
to the phenomenon
strain can be observed
being concentrated in theinindividual
Figure
10 aslayers.
CCF variation
Thisinphenomenon
graph noise for canthebetested
observedspecimen.
in Figure 10 as variation in graph noise for
the tested specimen.
To indicate the level of agreement between FE and test results, maximum principal
strains at the centre of the shaft and the strain concentration on the neck were compared
for similar shaft strain levels, and the strain concentration factor was calculated as a ratio
of neck to shaft strains. The corresponding values are given in Table 2.
 Materials 2022,15,
Materials2022, 15,1820
x FOR PEER REVIEW 11 11
ofof1514

Materials 2022, 15, x FOR PEER REVIEW 12 of 15

Table 3. Comparison of FE maximum principal stresses and Markforged data.

Value FE Max. Principal Stress CCF Strength

Absolute 734 MPa 800 MPa [32]
Relative 0.92 1.00

The strain at high load, shown in Figure 7 from nonlinear FEA, shows that compared
to lower loads, e.g., in Figure 8, the maximum strain at the eye shifts slightly towards the
shaft end. The fractured specimens in Figure 6 show the fracture line at the eye in a similar
position. The fracture initiation of the printed part coincides with the numerically highest
loaded area, giving confidence in the FEA.
Figure 10 shows the applied force on the test rig over the DIC-obtained maximum
principal strain in the shaft of the lug. For comparison, FEA-obtained shaft strain over
applied force is included. The data for tensile strength and a linear approximation of the
part stiffness are presented in Table 4. There is little variability in the strength, and the
fracture initiates at the location of maximum stress for all tests. The stiffness is similar for
all tested
Figure
Figure 9.9.lugs,
Majoraveraging
Major principal at 4.59
principal strain
strain με·N−1lug
in tested
in tested , and
lug No.agrees
No. 2.
2. well with the FEA results, 4.46 με·N−1.

To indicate the level of agreement between FE and test results, maximum principal
strains at the centre of the shaft and the strain concentration on the neck were compared
for similar shaft strain levels, and the strain concentration factor was calculated as a ratio
of neck to shaft strains. The corresponding values are given in Table 2.
In addition to the strain levels, the stiffness and strength of the lug were also
evaluated. As was shown earlier, the fracture occurred at the lug eye, perpendicular to the
load, and unfortunately, the DIC measurement could not capture this area because of the
rig setup. To estimate the stresses in this area of the part at failure, FE results were used.
Maximum principal stresses σmax,FE in the lug head were obtained for a load equivalent to
the average maximum force from experimental tests and are reported in Table 3. Because
the 2D FEA reports smeared results only, the stress values from FEA were corrected to
reflect that only every other layer of the material in the lug head is CCF-reinforced. This
value can be compared to the MEX composite coupons tested according to ASTM D3039
[31], and data derived from the supplier for fibre strength of the CCF material gives good
agreement.

Figure
Figure10.
10.Force
Forceover
overmaximum
maximumprincipal
principalstrains
strainsfor
forthe
thelug
lugshaft
shaftfrom
fromtesting.
testing.

Table 4. Strength and stiffness of the lugs from testing.

Tested Lugs Strength (kN) Linear Stiffness (με·

N−1)
Lug 2 11.81 4.93
Lug 3 11.32 4.77
Materials 2022, 15, 1820 12 of 14

In addition to the strain levels, the stiffness and strength of the lug were also evaluated.
As was shown earlier, the fracture occurred at the lug eye, perpendicular to the load, and
unfortunately, the DIC measurement could not capture this area because of the rig setup.
To estimate the stresses in this area of the part at failure, FE results were used. Maximum
principal stresses σmax,FE in the lug head were obtained for a load equivalent to the average
maximum force from experimental tests and are reported in Table 3. Because the 2D FEA
reports smeared results only, the stress values from FEA were corrected to reflect that only
every other layer of the material in the lug head is CCF-reinforced. This value can be
compared to the MEX composite coupons tested according to ASTM D3039 [31], and data
derived from the supplier for fibre strength of the CCF material gives good agreement.

Table 3. Comparison of FE maximum principal stresses and Markforged data.

Value FE Max. Principal Stress CCF Strength

Absolute 734 MPa 800 MPa [32]
Relative 0.92 1.00

The strain at high load, shown in Figure 7 from nonlinear FEA, shows that compared
to lower loads, e.g., in Figure 8, the maximum strain at the eye shifts slightly towards the
shaft end. The fractured specimens in Figure 6 show the fracture line at the eye in a similar
position. The fracture initiation of the printed part coincides with the numerically highest
loaded area, giving confidence in the FEA.
Figure 10 shows the applied force on the test rig over the DIC-obtained maximum
principal strain in the shaft of the lug. For comparison, FEA-obtained shaft strain over
applied force is included. The data for tensile strength and a linear approximation of the
part stiffness are presented in Table 4. There is little variability in the strength, and the
fracture initiates at the location of maximum stress for all tests. The stiffness is similar for
all tested lugs, averaging at 4.59 µε·N−1 , and agrees well with the FEA results, 4.46 µε·N−1 .

Table 4. Strength and stiffness of the lugs from testing.

Tested Lugs Strength (kN) Linear Stiffness (µε·N−1 )

Lug 2 11.81 4.93
Lug 3 11.32 4.77
Lug 4 11.32 4.64
Lug 5 11.66 4.03
Average 11.53 4.59
FEA - 4.46

Generally, the results obtained for the AM lug show that the FEA-assisted design of
fibre trajectories can predict the mechanical part properties well. The part stiffness and
strain distribution as well as the failure location and strength could be predetermined to
a good degree. The methodology can be easily applied to different AM structures, and
it is reasonable to believe that equally good results are achievable in terms of structural
stiffness and strain distribution. For part failure determination under more complex loading
conditions, a suitable multiaxial composite failure criterion has to be employed.

4. Conclusions
The potential of using additively manufactured CCF parts as load-carrying lugs has
been presented in this study. The part design was based on FEA and gave load-defined
fibre trajectories. Experimental tests with the printed parts were performed.
The main finding of this investigation is that FE analysis data and experimental
test results agreed well in terms of stiffness and strain. The FE-calculated part stiffness
was higher by a factor of 1.03, and the strain distributions were very similar. The strain
concentration around the neck differed by only 1%. Therefore, the potential of using
Materials 2022, 15, 1820 13 of 14

CAE-aided fibre steering for the design of additively manufactured CCF parts is given.
A further conclusion is that, potentially because there is some inherent variability in the
manufacturing process and in the material, a strength reduction of 8% in the manufactured
lugs compared to the manufacturer material data is yielded. As an application guideline, a
correction factor should be considered in engineering design to compensate for this.
In the present part, a viable fibre layup in line with the principal axis was obtained in
one analysis step. In future work, including analysis of more complicated part geometries,
the analysis may have to be performed in optimisation loops until the obtained principal
material axis deviates less than a certain set value from the obtained principal stress or
strain axis. Furthermore, limits on fibre orientation change over length, i.e., a minimum
radius of the fibre bundle, may have to be considered when performing more extensive
optimisation on complex geometries. Lastly, geometry optimisation was not part of the
objectives of the present study but may be a future addition to the method.

Supplementary Materials: The following supporting information can be downloaded at: https:
//www.mdpi.com/article/10.3390/ma15051820/s1, Figure S1: Prirevo CCF filament used in the
manufacturing of the lugs; Video S1: Video of CT Scan of fractured specimen, front view; Video S2:
Video of CT Scan of fractured specimen, side view.
Author Contributions: C.S. and S.S. conceptualised the research work, conducted experimental
design and data analysis and wrote the original draft. G.S. supervised and provided comments for im-
proving the manuscript. All authors have read and agreed to the published version of the manuscript.
Funding: This research work has been performed as part of the 3D CFRP (grant No. 859832). Open
Access Funding by the University of Linz.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Further data is available on request from the authors.
Acknowledgments: The Researchers acknowledge the financial support by the EU funded network
M-Era.Net, European Regional Development Fund (ERDF) through IWB 2014–2020, the BMVIT
(Austrian Ministry for Transport, Innovation, and Technology); the FFG (Austrian Research Promotion
Agency), the federal state of Upper Austria; and the RCL (Research Council of Lithuania) and FASIE
(Foundation for Assistance to Small Innovative Enterprises, Russia). Special thanks to Julia Maurer for
performing CT scans and the team Wood K plus for the operational assistance and fruitful discussions.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Fallon, J.J.; McKnight, S.H.; Bortner, M.J. Highly loaded fiber filled polymers for material extrusion: A review of current
understanding. Addit. Manuf. 2019, 30, 100810. [CrossRef]
2. De Toro, E.V.; Sobrino, J.C.; Martínez, A.M.; Eguía, V.M. Analysis of the influence of the variables of the Fused Deposition
Modeling (FDM) process on the mechanical properties of a carbon fiber-reinforced polyamide. Procedia Manuf. 2019, 41, 731–738.
[CrossRef]
3. Savandaiah, C.; Maurer, J.; Gall, M.; Haider, A.; Steinbichler, G.; Sapkota, J. Impact of processing conditions and sizing on the
thermomechanical and morphological properties of polypropylene/carbon fiber composites fabricated by material extrusion
additive manufacturing. J. Appl. Polym. Sci. 2021, 138, 50243. [CrossRef]
4. Savandaiah, C.; Plank, B.; Maurer, J.; Lesslhumer, J.; Steinbichler, G. Comparative Study of Filled and Unfilled Polylactic Acid
Produced via Injection Molding and 3D Printing. In Proceedings of the Plastic Technology Conference: SPE ANTEC® , Online,
10–21 May 2021.
5. Spoerk, M.; Savandaiah, C.; Arbeiter, F.; Traxler, G.; Cardon, L.; Holzer, C.; Sapkota, J. Anisotropic properties of oriented short
carbon fibre filled polypropylene parts fabricated by extrusion-based additive manufacturing. Compos. Part A Appl. Sci. Manuf.
2018, 113, 95–104. [CrossRef]
6. Ivey, M.; Melenka, G.W.; Carey, J.P.; Ayranci, C. Characterizing short-fiber-reinforced composites produced using additive
manufacturing. Adv. Manuf. Polym. Compos. Sci. 2017, 3, 81–91. [CrossRef]
7. Kabir, S.; Mathur, K.; Seyam, A.-F. Maximizing the Performance of 3D Printed Fiber-Reinforced Composites. J. Compos. Sci. 2021,
5, 136. [CrossRef]
Materials 2022, 15, 1820 14 of 14

8. Brooks, H.; Molony, S. Design and evaluation of additively manufactured parts with three dimensional continuous fibre
reinforcement. Mater. Des. 2016, 90, 276–283. [CrossRef]
9. Dul, S.; Fambri, L.; Pegoretti, A. High-Performance Polyamide/Carbon Fiber Composites for Fused Filament Fabrication:
Mechanical and Functional Performances. J. Mater. Eng. Perform. 2021, 30, 5066–5085. [CrossRef]
10. Consul, P.; Beuerlein, K.-U.; Luzha, G.; Drechsler, K. Effect of Extrusion Parameters on Short Fiber Alignment in Fused Filament
Fabrication. Polymers 2021, 13, 2443. [CrossRef]
11. Mulholland, T.; Goris, S.; Boxleitner, J.; Osswald, T.A.; Rudolph, N. Process-Induced Fiber Orientation in Fused Filament
Fabrication. J. Compos. Sci. 2018, 2, 45. [CrossRef]
12. Boros, R.; Rajamani, P.K.; Kovacs, J.G. Combination of 3D printing and injection molding: Overmolding and overprinting. Express
Polym. Lett. 2019, 13, 889–897. [CrossRef]
13. De Toro, E.V.; Sobrino, J.C.; Martínez, A.M.; Eguía, V.M.; Pérez, J.A. Investigation of a Short Carbon Fibre-Reinforced Polyamide
and Comparison of Two Manufacturing Processes: Fused Deposition Modelling (FDM) and Polymer Injection Moulding (PIM).
Materials 2020, 13, 672. [CrossRef]
14. Liu, G.; Xiong, Y.; Zhou, L. Additive manufacturing of continuous fiber reinforced polymer composites: Design opportunities
and novel applications. Compos. Commun. 2021, 27, 100907. [CrossRef]
15. Zhuo, P.; Li, S.; Ashcroft, I.A.; Jones, A.I. Material extrusion additive manufacturing of continuous fibre reinforced polymer
matrix composites: A review and outlook. Compos. Part B Eng. 2021, 224, 109143. [CrossRef]
16. Azarov, A.V.; Antonov, F.K.; Golubev, M.V.; Khaziev, A.R.; Ushanov, S.A. Composite 3D printing for the small size unmanned
aerial vehicle structure. Compos. Part B Eng. 2019, 169, 157–163. [CrossRef]
17. Borowski, A.; Vogel, C.; Behnisch, T.; Geske, V.; Gude, M.; Modler, N. Additive Manufacturing-Based In Situ Consolidation of
Continuous Carbon Fibre-Reinforced Polycarbonate. Materials 2021, 14, 2450. [CrossRef]
18. Dickson, A.N.; Barry, J.N.; McDonnell, K.A.; Dowling, D.P. Fabrication of continuous carbon, glass and Kevlar fibre reinforced
polymer composites using additive manufacturing. Addit. Manuf. 2017, 16, 146–152. [CrossRef]
19. Zhu, Y.; Liu, J.; Liu, D.; Xu, H.; Yan, C.; Huang, B.; Hui, D. Fiber path optimization based on a family of curves in composite
laminate with a center hole. Compos. Part B Eng. 2017, 111, 91–102. [CrossRef]
20. Sugiyama, K.; Matsuzaki, R.; Malakhov, A.V.; Polilov, A.N.; Ueda, M.; Todoroki, A.; Hirano, Y. 3D printing of optimized
composites with variable fiber volume fraction and stiffness using continuous fiber. Compos. Sci. Technol. 2020, 186, 107905.
[CrossRef]
21. Ferreira, R.T.L.; Ashcroft, I.A.; Li, S.; Zhuo, P. Optimisation of Fibre-Paths in Composites Produced by Additive Manufacturing.
In EngOpt 2018 Proceedings of the 6th International Conference on Engineering Optimization, Lisbon, Portugal, 17–19 September 2018;
Rodrigues, H.C., Herskovits, J., Mota Soares, C.M., Araújo, A.L., Guedes, J.M., Folgado, J.O., Moleiro, F., Madeira, J.F.A., Eds.;
Springer International Publishing: Cham, Switzerland, 2019; pp. 1083–1094, ISBN 978-3-319-97772-0.
22. Ghiasi, H.; Fayazbakhsh, K.; Pasini, D.; Lessard, L. Optimum stacking sequence design of composite materials Part II: Variable
stiffness design. Compos. Struct. 2010, 93, 1–13. [CrossRef]
23. Ghiasi, H.; Pasini, D.; Lessard, L. Optimum stacking sequence design of composite materials Part I: Constant stiffness design.
Compos. Struct. 2009, 90, 1–11. [CrossRef]
24. Papapetrou, V.S.; Patel, C.; Tamijani, A.Y. Stiffness-based optimization framework for the topology and fiber paths of continuous
fiber composites. Compos. Part B Eng. 2020, 183, 107681. [CrossRef]
25. Setoodeh, S.; Abdalla, M.M.; Gürdal, Z. Design of variable–stiffness laminates using lamination parameters. Compos. Part B Eng.
2006, 37, 301–309. [CrossRef]
26. Chacón, J.M.; Caminero, M.A.; Núñez, P.J.; García-Plaza, E.; García-Moreno, I.; Reverte, J.M. Additive manufacturing of
continuous fibre reinforced thermoplastic composites using fused deposition modelling: Effect of process parameters on
mechanical properties. Compos. Sci. Technol. 2019, 181, 107688. [CrossRef]
27. Schaberger, M. Damage Detection in Thin-Walled Structures with Strain Measurements along Zero-Strain Trajectories. Master’s
Thesis, Institute of Structural Lightweight Design, Johannes Kepler University, Linz, Austria, 2016.
28. Sanei, S.H.R.; Popescu, D. 3D-Printed Carbon Fiber Reinforced Polymer Composites: A Systematic Review. J. Compos. Sci. 2020, 4,
98. [CrossRef]
29. Shiratori, H.; Todoroki, A.; Ueda, M.; Matsuzaki, R.; Hirano, Y. Mechanism of folding a fiber bundle in the curved section of 3D
printed carbon fiber reinforced plastics. Adv. Compos. Mater. 2020, 29, 247–257. [CrossRef]
30. Shiratori, H.; Todoroki, A.; Ueda, M.; Matsuzaki, R.; Hirano, Y. Compressive strength degradation of the curved sections of
3D-printed continuous carbon fiber composite. Compos. Part A Appl. Sci. Manuf. 2021, 142, 106244. [CrossRef]
31. ASTM D3039. Standard Test Method for Tensile Properties of Polymer Matrix Composite Materials. Available online: https:
//www.astm.org/d3039_d3039m-17.html (accessed on 26 February 2022).
32. Markforged-CCF Material Datasheet. Available online: https://markforged.com/materials/continuous-fibers/continuous-
carbon-fiber (accessed on 29 December 2021).